INDEX

Explanations

phrases indicating an abstract or conceptual understanding or interpretation of a situation or concept

oai_token-act-pair · gpt-3.5-turbo

phrases that reference the concept of "sense" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 helicop

-0.74

sites

-0.73

onial

-0.71

imer

-0.71

annis

-0.69

 Crimson

-0.68

 Byrd

-0.67

 tatt

-0.66

etsk

-0.64

 Vanderbilt

-0.63

POSITIVE LOGITS

ombat

0.74

 sense

0.73

Ëľ

0.71

 compulsion

0.70

 Sense

0.67

ÙĨ

0.67

 Ø§ÙĦ

0.66

ively

0.66

forward

0.64

 meaning

0.64

Activations Density 0.012%

phrases indicating an abstract or conceptual understanding or interpretation of a situation or concept

phrases that reference the concept of "sense" in various contexts

No Comments

No Known Activations